Changes in Chemical Properties of Sandy Duplex Soils in 11 Paddocks over 21 Years in the Low Rainfall Cropping Zone of Southwestern Australia

Soil testing was conducted during 1985-2005 in 11 paddocks on sandy duplex soils on Newdegate Research Station, average annual rainfall of 377 mm, with about 70% falling in the May-October growing season, in the Mediterranean-type climate of southwestern Australia. The study was undertaken to determine lime and fertilizer requirements of eight crop species grown in rotation with one another (one crop each year in the typical May-October growing season, comprising wheat, Triticum aestivum L.; barley, Hordeum vulgare L.; oats, Avena sativa L.; lupin, Lupinus angustifolius L.; canola, Brassica napus L.; chickpea, Cicer arietinum L.; field pea, Pisum sativum L.; and subterranean clover-based pasture, Trifolium subterraneum L. All crops were sown using no-till. The study demonstrated that plant testing was required in conjunction with soil testing to confirm decisions based on soil testing and to assess management decisions for elements not covered by soil testing. Pasture dry-matter production seldom exceeds 2 t ha(-1) during the growing season in the region, but clover pasture is valued as a break crop for diseases and pests of grain crops and to facilitate control of herbicide-resistant weeds for cropping. Pastures had negligible impact on soil-test values. By contrast, grain crops typically produce more dry matter than pasture (4-8 t ha(-1)) and consistently significantly resulted in soil pH, soil-test potassium (K), and organic carbon (C) of soil decreasing through time. Fertilizer phosphorus (P) was not applied to pasture but was applied while sowing most grain crops from 1985 to 1996, a common practice at the time, and soil-test P significantly increased through time in these years. Thereafter fertilizer P was only applied when soil-test P was less than the critical value for that soil and grain crop species resulting, in little P being applied in these years, and soil-test P significantly declined through time. Plant testing indicated P was adequate when soil testing indicated no fertilizer P was required. The soils only started to become K deficient in the mid-1990s because of the removal of indigenous soil K in grain, and fertilizer K was applied when soil-test K was less than the 50 mg kg(-1) critical value determined for wheat and canola. Plant testing indicated K was adequate when soil testing indicated no fertilizer K was required, and it indicated K was adequate after fertilizer K was applied, showing K levels applied were adequate for grain production. Plant testing indicated nitrogen (N), sulfur (S), calcium (Ca), magnesium (Mg), copper (Cu), zinc (Zn), manganese (Mn), iron (Fe), and boron (B) were adequate for grain production. Electrical conductivity (EC) of soil was very variable but EC values indicated soil salinity was unlikely to reduce grain yields of all the crop species grown. We conclude soil testing for pH is reliable for indicating paddocks requiring lime to ameliorate soil acidity and to monitor progress of liming. Soil testing proved reliable for determining when fertilizer P and K needed to be applied. Research has shown that for the low rainfall cropping areas of southwestern Australia laboratories need to measure and report soil pH, soil-test P, and soil-test K every 1-3 years and the P-buffering index (estimating P sorption of soil), organic C, and electrical conductivity every 3-5 years.